EventPlanGenie

An Agentic AI-powered Multi-Agent Event Planning System using LangGraph, Streamlit & FastAPI

Built with LangGraph + LangChain | Modular AI Agents | Local-First Architecture

---

Overview

EventPlanGenie is a smart event planning assistant powered by multiple collaborating AI agents. It uses local LLMs and a modular architecture to automate creative and logistical planning task.

Plan smarter, faster, and locally — no cloud dependencies.

---

Features

• Fully local execution using LLMs via Ollama (phi3)
• Modular multi-agent system (Idea, Vendor, Scheduler, Invitation, Reviewer)
• Markdown and PDF export
• Email-ready event invitations
• Agent coordination via LangGraph protocol
• MCP (Model-Context-Protocol) design pattern

---

Tech Stack

Layer	Tech Used
Frontend	Streamlit
Backend	FastAPI, LangGraph, LangChain
LLMs	phi3 via Ollama
Export	markdown, reportlab
Emailing	smtplib, email

---

Project Structure

eventplangenie/
├── agents/                  # All AI agents (Idea, Vendor, Scheduler, etc.)
│   ├── base_agent.py
│   ├── idea_agent.py
│   ├── vendor_agent.py
│   ├── scheduler_agent.py
│   ├── invitation_agent.py
│   ├── reviewer_agent.py
│   └── email_agent.py
│
├── coordinator/            # LangGraph protocol graph (MCP)
│   └── graph.py
│
├── backend/                # FastAPI backend logic
│   └── main.py
│
├── frontend/               # Streamlit user interface
│   ├── streamlit_app.py
│   └── streamlit_app2.py
│
├── utils/                  # Helper utilities
│   ├── markdown_formatter.py
│   ├── pdf_helper.py
│   └── ics_generator.py    # (optional – remove if not used)
│
├── .env.example            #  Sample environment configuration
├── requirements.txt        #  Python dependencies
├── .gitignore              #  Git ignore rules
└── README.md               #  Project documentation

---

Agent Roles

Agent	Role
IdeaAgent	Suggests creative event ideas
VendorAgent	Finds vendors based on event type/location
SchedulerAgent	Plans the timeline and event schedule
InvitationAgent	Generates a well-written invite (Markdown + PDF)
ReviewerAgent	Reviews final output and suggests improvements

All agents interact in LangGraph, forming a real-time decision flow.

---

How to Run

1. Clone and Setup

git clone https://github.com/mrenikareddy06/EventPlanGenie.git
cd EventPlanGenie
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Configure Environment

cp .env.example .env
Update with your email config (if using email invites).

3. Start Backend

uvicorn backend.main:app --reload

4. Launch Frontend

streamlit run frontend/streamlit_app.py

Output Formats

• PDF: Downloadable summary
• Email: Send invitation directly

Architecture (Visualized)

EventPlanGenie-v3